In the paper you mention that you train the network with RD maps of size 512x128. In the last section of the paper you mention that you evaluate the network on the data with different number of chirp values like 16, 32, 64. My question is - how do you fit the smaller input (512x{16, 32, 64}) during the evaluation into the architecture expecting 512x128 shape of the data?
Hello,
In the paper you mention that you train the network with RD maps of size 512x128. In the last section of the paper you mention that you evaluate the network on the data with different number of chirp values like 16, 32, 64. My question is - how do you fit the smaller input (512x{16, 32, 64}) during the evaluation into the architecture expecting 512x128 shape of the data?
Thanks for the info!